upa - home page JUS - Journal of usability studies
An international peer-reviewed journal

Card Sort Analysis Best Practices

Carol Righi, Janice James, Michael Beasley, Donald L. Day, Jean E. Fox, Jennifer Gieber, Chris Howe, and Laconya Ruby

Journal of Usability Studies, Volume 8, Issue 3, May 2013, pp. 69 - 89

Article Contents


Information architecture (IA) is the practice of effectively organizing, structuring, and labeling the content of a website or application. IA is used to create the navigation structure of a website, including the categories of content and what they are named. Card sorting is a popular user-centered design method that helps us derive an effective IA. At first glance, card sorting seems to be a deceptively easy task: Present representative users with content or functionality slated to be included in a website or application, and ask them to sort it into logical groupings. What better way to create an IA than to get the users themselves to create it? There is little to be argued with when looking at card sorting this way.

In recent years, the process of conducting card sorts has gone from handing a stack of 3 x 5 index cards to users to creating an online study, emailing a link to users, and sitting back and collecting the data. Again: Deceptively easy.

However, with card sorting, as the saying goes, the devil is in the details. Once you’ve collected the data, how do you make sense of it? How do you discern and assess the groupings the users created? How do you compile results from many users? How do you reconcile disparate groupings? How do you determine the best labels for the groupings? And, how do you do all of this in a way that, as much as possible, is backed by solid, objective criteria?

Fortunately, the tools available today to collect the data include electronic data visualization and analysis routines to help analyze the data. Online dendrograms and various item/category matrixes quickly and easily provide the practitioner with a variety of views of the data. The use of these tools and the interpretation of the resulting data, though, are far less straightforward.

To address the vexing problem of card sort data analysis, a group of User Experience (UX) practitioners convened at the 2011 Usability Professionals’ Association conference in Atlanta to address and formulate a set of best practices for analyzing card sort data. This paper details those practices.

This paper does not address the question of which tools to use or the relative merits of variations of card sorting. And, because the focus of this paper is on the analysis of the collected data, we also do not discuss the planning and creation of a card sorting study. Most of what is addressed will refer to various automated card sorting tools, but this paper will largely be tool-agnostic. For the purposes of illustrating the various tools and methods, this paper will, however, reference a sample card sort study for a coffee house’s website that we performed using the WebSort tool (http://uxpunk.com/websort/; URL as of 10-21-12).

This paper presents the approach in a more-or-less sequential order of when we recommend each step be implemented. However, depending on the situation, you may change the order or eliminate some steps entirely. Although there are many informal methods for analyzing card sort data, the approach described here is more objective and data-based compared to a simple “eye balling” of the sort results.

Even though these guidelines are presented as a sequence, card sort data analysis, like many research practices, is typically a highly iterative process. You will likely wish to revisit earlier steps in the sequence to get different views of the data or to test alternate hypotheses regarding the IA.

Finally, this set of guidelines is just that: a guide for approaching a sometimes complex but much-needed technique in the arsenal of the UX professional.


Previous | Next